Magnetic thin films and multilayers play a key role in various types of magnetic storage devices such as a magnetic hard disk (HDD) drive, Magnetic Random Access Memory (MRAM), spin torque oscillator (STO), and magnetic domain wall devices. In order to develop and optimize such devices, monitoring and characterization of magnetic thin film stacks are necessary. A variety of different magnetic characterization techniques must be used to determine all the essential magnetic parameters such as crystalline anisotropy, surface or interface anisotropy, magnetization saturation (Ms), damping constant (α), gyromagnetic ratio (γ), inhomogeneous broadening, resistance x area product (RA), and magnetoresistive ratio (MR).
FMR is a well-established method of measuring anisotropy fields, as well as the gyromagnetic ratio γ, and the damping constant α of magnetic films and multilayers in extended unpatterned films or over an area comprising a large array of sub-micron patterned structures. The resonance frequency fR of a ferromagnetic film is given by the so-called Kittel formula shown in equation (1) below where HR is the resonance field applied perpendicular to the plane of the film, HK is the effective anisotropy field which includes structural, surface, and magnetostatic contributions, and γ is the gyromagnetic ratio.2πfR=γ(HR+HK)  (Eq. 1)
A FMR experiment is performed by probing the magnetic system (thin film, multilayer stack, or structured device) with a combination of microwave excitation and a quasi-static magnetic field. FMR data is obtained by either sweeping the magnetic field at a constant microwave frequency, or by sweeping the frequency at a constant field. When the ferromagnetic resonance condition is achieved, it may be detected by an enhanced absorption of the microwave (RF signal) by the ferromagnetic sample. Thus, resonance (FMR) conditions are defined with pairs of magnetic field and microwave frequency values (HR, fR).
There are several ways of submitting a ferromagnetic sample to microwave excitation. Historically, FMR experimental conditions employed tubular waveguides, and samples were placed in a resonant cavity between poles of an electromagnet. More recently, new methods have been developed that are well suited to analyze film shaped samples. In particular, the wafer under test (WUT) is placed in contact with a non-magnetic waveguide transmission line (WGTL) that may be in the form of a grounded coplanar waveguide (GCPWG), coplanar waveguide (CPWG), co-axial waveguide (CWG), stripline (SL), or a microstrip (MS). The power transmitted or reflected by the WGTL is monitored as a function of the applied magnetic field and microwave frequency.
Referring to FIG. 1a, a schematic depiction is shown where output voltages are plotted as a function of a variable magnetic field at constant microwave frequency (f) using five different values (f1-f5) of microwave frequency. The center and width of the Lorentzian peaks is extracted from the data as a function of the excitation frequency. As mentioned previously, the center field is the resonance field (HR), which is related to the excitation frequency following the Kittel formula that is rewritten in a slightly different form in equation (2) below where h is the Planck constant and μB is the Bohr magneton.HR(f)=[h/(γ×μB)]×f−HK  (Eq. 2)
The variation of HR with microwave frequency is shown in FIG. 1b where each of the points along curve 21 is derived from one of the Lorentzian shaped peaks Hr1-Hr5 in FIG. 1a. As indicated by equation (2), the extrapolation of the data to f=0 gives the value of the effective anisotropy field HK.
The linewidth L of the resonance peak is the width at half amplitude ΔH of the resonance peak and is related to dissipative processes involved in magnetization dynamics.
The linewidth depends on the excitation frequency and the dimensionless Gilbert damping constant according to equation (3) below where L0 is an inhomogeneous broadening. By fitting HR and L with respect to the excitation frequency fR, HK as well as α and γ may be derived.L(f)=(2hα/γ×μB)f+L0  (Eq. 3)
A network analyzer for detecting FMR in thin CoFe and CoFeB films on a coplanar waveguide is described by C. Bilzer et al. in “Vector network analyzer ferromagnetic resonance of thin films on coplanar waveguides: Comparison of different evaluation methods” in J. of Applied Physics, 101, 074505 (2007), and in “Open-Circuit One-Port Network Analyzer Ferromagnetic Resonance” in IEEE Trans. Magn., Vol. 44, No. 11, p. 3265 (2008). In these experiments, the planar WGTL is typically attached to radiofrequency (RF) connectors by microwave electrical probes and placed between the poles of an electromagnet. Thus, given the size of the WGTL (about 5 mm long), and the size of the gap of typical electromagnets, only small size samples (normally<1 inch in diameter) can be measured. Accordingly, wafers typically used in the microelectronics industry (having diameters of 6, 8, 12 inches or more) can only be measured with this FMR technique if they are cut into small coupons.
FIG. 2 is reproduced from “Microwave susceptibility of thin ferromagnetic films: metrology and insight into magnetization dynamics”, Claus Bilzer, Ph. D. report, Universite Paris Sud—Paris XI, 2007) and depicts a conventional FMR system. Vector network analyzer (VNA) 10 is connected from one port through a first coaxial cable 1 to a first microwave probe portion 2a that is attached to coplanar waveguide (CPWG) 6. A top surface of the CPWG adjoins a magnetic film 4 mounted on a substrate. When a microwave frequency from probe 2a and an external magnetic field 3 are applied in an x-axis direction across the magnetic film, an output signal in a transmission mode passes into a second microwave probe portion 2b and then through a second coaxial cable 5 before returning to the VNA at a second port. Magnetic film size b is typically restricted to 1 inch or less, which means the magnetic film sample 4 must be cut from a whole wafer.
Since conventional FMR techniques are destructive, time consuming, and limited to measuring large structures having a cross-sectional size substantially greater than 1 mm, they are undesirable to an extent that prevents wide acceptance of FMR as a characterization tool in the magnetic data storage industry. An improved FMR measurement system and technique is needed that enables fully automated measurements on whole wafers for faster throughput and lower cost. Preferably, the FMR system may be constructed from commercially available parts. Also, the improved FMR technique should be able to measure magnetic properties in portions of films having a diameter of less than 1 mm to enable smaller test structures to be measured than is possible in the prior art.